bookbot/sherpa-onnx-ort-zipformer-streaming-robust-sw-v4

Sherpa-onnx Pruned Stateless Zipformer RNN-T Streaming Robust SW v4

Sherpa-onnx Pruned Stateless Zipformer RNN-T Streaming Robust SW v4 is an automatic speech recognition model trained on the following datasets:

• ALFFA Swahili
• FLEURS Swahili
• Common Voice 16.1 Swahili
• Bookbot Swahili v2 Filtered
• Bookbot Swahili Noisy v1

Instead of being trained to predict sequences of words, this model was trained to predict sequence of phonemes, e.g. ["w", "ɑ", "ʃ", "i", "ɑ"]. Therefore, the model's vocabulary contains the different IPA phonemes found in gruut.

This model was converted from the TorchScript version of Pruned Stateless Zipformer RNN-T Streaming Robust SW to ONNX format.

Converting from TorchScript

Refer to the official instructions for conversion to ONNX.

Frameworks

• k2
• icefall
• lhotse
• sherpa-onnx
• onnx
• onnxruntime